HUB LOCATION AND NETWORK DESIGN: MODEL FORMULATION AND NUMERICAL ANALYSES Guoqiang Shen   A general discrete hub location and network design model is developed that accounts for fixed, capacity, congestion, and reliability costs, characterized by both economies and diseconomies of scale, selects hubs and links, determines their capacities, and assigns O-D flows over paths, while minimizing total system costs. Initially formulated as a mixed-integer non-linear program (MINLP), the model is transformed into a mixed-integer linear program (MILP) by linearizing the capacity and congestion cost functions. The MILP model is solved for a small-scale network with hypothetical data. Extensive numerical experiments, including a numerical analysis and a sensitivity analysis, demonstrate that there are significant trade-offs between the different link and hub costs, and that the optimal networks produced by the model can take the form of a complete point-to-point (P-P) network, a hub-and-spoke (H-S) network, or a minimum spanning tree (MST), underscoring the fact that the H-S structure does not always guarantee the minimum total system cost. An extensive reviewof the hub literature shows that, by adding more restrictive constraints, the existing hub models can be derived from the general hub location and network design model.